Reflection and summarization of setting up Hadoop and HBase cluster

These days at the teacher's request, in three machines to build Hadoop environment and http://www.aliyun.com/zixun/aggregation/13713.html ">hbase environment, which encountered a lot of problems, until today the basic operation of success." Configuration details are not discussed here, but the issues needing attention are listed for reference.

Three machines time synchronization, firewall (iptables) shutdown

Set hostname. For Debian, modify the/etc/hostname; for CentOS, modify/etc/sysconfig/network.

Sets the Hosts file. Remove the 127.0.0.1 host name and 127.0.1.1 hostname (if any); 127.0.0.1 localhost can be retained. Add the IP address and host name of each machine.

Set SSH no password access between each machine. You must ensure that no password access is successful between 22 machines.

Adjust the limits of the machine. See here for details (official website)

To set the profile hadoop-env.sh in Hadoop, you need to set the Java_home

HBase configuration file hbase-env.sh, you need to set the java_home;

Core-site.xml, Hdfs-site.xml, hbase-site.xml core file configuration. The related configuration is visible here.

Deploy Hadoop and HBase to each node. Note the consistency of the profiles in each node, as well as the validity (such as the Java_home path, the local Java path required if the machine is different, and the local storage path for HDFs and zookeeper)

Start HDFs First, then start HBase.

After this setup, there are warnings in hbase-hadoop-master-psydebian.out and when creating tables:

is a jar pack duplication issue. But if I delete one of the Hadoop or hbase, Hadoop or hbase will complain when I start or stop. The package duplication problem is not currently resolved, but it does not seem to affect the use of hbase such as creating tables.